Blind source separation for impact force induced by piston slaps in diesel engines

The diesel engine vibration induced by piston impact force (also known as slaps) presents us very important information about piston-and-cylinder clearance and engine performance. However it is notorious difficult for making an accurate measurement for the piston slaps due to the enormous interferences, reverberation, and complicated vibration responses. This paper presents an application of blind source separation technique to solve the problem. The method is based on one novel algorithm which uses blind source de-convolution algorithm. It can overcome the conventional blind source separation calculation drawbacks like local minimization value problems occurred with the fixed step response algorithm. The simulation and experiment results have indicated it can recover the corrupted signals and obtain the accurate piston slap signals at the measurement point.